rapidsai/cugraph

Health v2 · maintenance only

Adoption not part of the score

2225 stars · 365 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-30, confidence not recorded

cuGraph is NVIDIA's RAPIDS collection of GPU-accelerated graph analytics libraries, offering Python, C, and C++ APIs for building graphs and running scalable graph algorithms like PageRank, centrality, community detection, and traversal. It integrates with cuDF/pandas DataFrames, NetworkX (via nx-cugraph), and PyTorch Geometric for graph neural network workflows.

Use cases

• run pagerank on large graphs on gpu
• accelerate networkx algorithms with zero code changes
• detect communities with louvain or leiden at scale
• compute shortest paths and centrality on massive graphs
• train graph neural networks with pytorch geometric on gpu
• multi-gpu distributed graph analytics
• find connected components in billion-edge graphs

When to choose

• you have NVIDIA GPUs and need fast graph algorithms on large graphs
• your NetworkX workflows are too slow and you want a drop-in backend
• you're doing GNN training and need GPU graph sampling and storage
• you want seamless integration with RAPIDS cuDF ETL pipelines

When to avoid

• you have no CUDA-capable GPU hardware
• your graphs are small enough that CPU libraries like NetworkX suffice
• you need algorithms not in cuGraph's supported list
• you need a pure-CPU or portable solution

Facets

library · maturity active

machine-learning data-science etl graphics graph-processing machine-learning data-science gpu-computing python cpp graph-analytics cuda rapids networkx-backend graph-neural-networks pagerank community-detection multi-gpu algorithms gpu linux docker

Member repositories

For agents

markdown · JSON · MCP: product_card(name="rapidsai/cugraph")